Here's a statement of the obvious: The opinions expressed here are those of the participants, not those of the Mutual Fund Observer. We cannot vouch for the accuracy or appropriateness of any of it, though we do encourage civility and good humor.
Thank you for the four disparate references to correlation coefficient perspectives.
Indeed, candidate investment option correlations are crucial when assembling a portfolio with reduce volatility (standard deviation). You are doing a worthwhile search and service for MFO members.
Unfortunately, the first three of your references are either self-serving or rather pedestrian. They make assertions without providing reliable data or references. In essence they simply offer opinions. In particular, the first article, instead of providing a reasoned critique of correlation methods, just pontificates while extolling the author’s own biases and investment worldview.
The fourth article does some original Monte Carlo analysis and warns of the dangers of stale long term data that anchors not current investment pair correlations. That is a real threat. There is always a balancing act when doing correlation work; the benefits of longer stale data statistically must be measured against the freshness of more limited but more pertinent current data. The tradeoff is not easy since correlation coefficients can vary considerably over time.
You can test the issue yourself and select an appropriate timeframe by doing an independent analysis. I discussed the matter in a late May posting titled “Diversity with Correlation Coefficients”. Here is the internal Link to my submittal:
Comments
Thank you for the four disparate references to correlation coefficient perspectives.
Indeed, candidate investment option correlations are crucial when assembling a portfolio with reduce volatility (standard deviation). You are doing a worthwhile search and service for MFO members.
Unfortunately, the first three of your references are either self-serving or rather pedestrian. They make assertions without providing reliable data or references. In essence they simply offer opinions. In particular, the first article, instead of providing a reasoned critique of correlation methods, just pontificates while extolling the author’s own biases and investment worldview.
The fourth article does some original Monte Carlo analysis and warns of the dangers of stale long term data that anchors not current investment pair correlations. That is a real threat. There is always a balancing act when doing correlation work; the benefits of longer stale data statistically must be measured against the freshness of more limited but more pertinent current data. The tradeoff is not easy since correlation coefficients can vary considerably over time.
You can test the issue yourself and select an appropriate timeframe by doing an independent analysis. I discussed the matter in a late May posting titled “Diversity with Correlation Coefficients”. Here is the internal Link to my submittal:
http://www.mutualfundobserver.com/discussions-3/#/discussion/6667/diversity-with-correlation-coefficients-ping-hogan
The referenced posting includes a Link to a very user friendly correlation coefficient calculator. It works. Please give it a try.
Thanks again for the fourth reference.
Best Wishes.